Dear Professor,
I am a postgraduate student. I have recently been reading your paper[A deep learning approach to wall-shear stress quantification: from numerical training to zero-shot experimental application] and found your research very interesting and inspiring. I am currently trying to reproduce part of your work for my own research.
However, I encountered some difficulties when downloading and preparing the dataset according to the instructions in the paper. Based on the description in your work, I expected the training data format to be (N, C, Z, Y, X). But the dataset I downloaded from the website[https://turbulence.idies.jhu.edu/] consists of files named sample_000000 to sample_009599, and the shape of each sample appears to be (128, 128, 3).
Therefore, I would like to ask whether any preprocessing steps were applied to the original dataset before training. If so, I would be very grateful if you could kindly let me know what preprocessing operations were performed.It would be even better if you could share the dataset files and Python preprocessing code you used for the experiments. Alternatively, if I may have made a mistake when selecting the download options or interpreting the dataset format, I would also sincerely appreciate your clarification.
Your guidance would be extremely helpful for my research. If I use your method and dataset setting in my work, I will of course cite your paper appropriately in my future publication.
Thank you very much for your time and help. I look forward to your reply.
Best regards,
Chiuvin Chen
Dear Professor,
I am a postgraduate student. I have recently been reading your paper[A deep learning approach to wall-shear stress quantification: from numerical training to zero-shot experimental application] and found your research very interesting and inspiring. I am currently trying to reproduce part of your work for my own research.
However, I encountered some difficulties when downloading and preparing the dataset according to the instructions in the paper. Based on the description in your work, I expected the training data format to be (N, C, Z, Y, X). But the dataset I downloaded from the website[https://turbulence.idies.jhu.edu/] consists of files named sample_000000 to sample_009599, and the shape of each sample appears to be (128, 128, 3).
Therefore, I would like to ask whether any preprocessing steps were applied to the original dataset before training. If so, I would be very grateful if you could kindly let me know what preprocessing operations were performed.It would be even better if you could share the dataset files and Python preprocessing code you used for the experiments. Alternatively, if I may have made a mistake when selecting the download options or interpreting the dataset format, I would also sincerely appreciate your clarification.
Your guidance would be extremely helpful for my research. If I use your method and dataset setting in my work, I will of course cite your paper appropriately in my future publication.
Thank you very much for your time and help. I look forward to your reply.
Best regards,
Chiuvin Chen